Journal of CO2 Utilization 68 (2023) 102378
Available online 30 December 2022
2212-9820/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Carbon molecular sieve membranes for selective CO 2 separation at elevated temperatures and pressures
A. Rahimalimamaghani
a, H.R. Godini
a,*, M. Mboussi
a, A. Pacheco Tanaka
a,b, M. Llosa Tanco
a,b, F. Gallucci
a,c,**aInorganic Membranes and Membrane Reactors, Sustainable Process Engineering, Chemical Engineering and Chemistry, Eindhoven University of Technology, Eindhoven, the Netherlands
bTECNALIA, Basque Research and Technology Alliance (BRTA), Mikeletegi Pasealekua 2, 20009, Donostia, San Sebastian, Spain
cEindhoven Institute for Renewable Energy Systems (EIRES), Eindhoven University of Technology, PO Box 513, Eindhoven 5600 MB, the Netherlands
A R T I C L E I N F O Keywords:
Carbon membranes CO2 separation Post-combustion capture Process intensification
Multi-components permeation tests
A B S T R A C T
The use of Carbon Molecular Sieve Membranes (CMSM) for selective CO2 separation from post-combustion CO2- rich streams from steel plant was experimentally evaluated and reported in this paper. Efficient CMSMs were developed for such application and their promising potentials in operating at elevated temperatures and pres- sures were experimentally demonstrated. The best performance in terms of flux as well as perm-selectivity, above the Robeson upper bound, was obtained using a CMSM developed with ethylenediamine in the dip-coating stage of the fabrication. In fact, adding ethylenediamine was proven to be particularly important in narrowing down the pore size distribution to ultra-micropore and establishing effective CO2 adsorption site over the membrane surface and the pores wall. It was shown that using a tailored CMSM with a precursor synthesized by co- polymerization of ethylenediamine with Novolac can improve the CO2/N2 ideal perm-selectivity from 33 to 97 at operational conditions of 200 ◦C and 20 bar.
1. Introduction
Reducing the anthropogenic carbon dioxide emissions has became an ever-pressing need with far reaching environmental, economic and so- cietal impacts [1]. Depending on the specific conditions and the re- quirements of each CO2 generating sector, proper technologies should be carefully chosen and implemented for the separation and possibly storage or utilization of CO2 [2]. Focusing on the steel production plants, these account for around 25 % of the emitted CO2 from industrial pro- cesses [3], being one of the main sector-contributors in the global carbon dioxide emission pushing it beyond 36 Giga tonne annually [4,5]. The relatively high CO2 content of the steel mill gases should be preferably separated at elevated temperatures to secure an energy-efficient sepa- ration process with a lower carbon footprint. 45 % carbon oxides content is a typical carbon-containing portion of a mill gas [6], accompanied with higher nitrogen content and smaller amounts of other components such as hydrogen. Therefore, two of the main challenges to be focused on in this case are the conditions of the treated off-gas in terms of its high
temperature and composition. There are not many practically feasible separation-technologies, which can address these challenges efficiently in terms of securing a simultaneous selective and energy efficient CO2
recovery from such hot multi-components CO2-rich gaseous mixtures [7]. Among them, membrane separation is regarded as an interesting technology for low-cost CO2 separation. Amongst the different possible membrane technologies, zeolite membranes can also theoretically be considered. However, they suffer stability problems at high temperature in presence of steam. Additionally, it is very challenging to reproducibly produce defect free zeolite membranes with high selectivity and low costs. Carbon Molecular Sieve Membranes (CMSMs) with proven effi- ciency in selective CO2-separation at elevated temperatures [8] is a promising option to zeolite membranes. CMSMs could surpass the Robeson’s upper bound limit for membrane perm-selectivity of poly- meric membranes [9,10] and yet do not counter the problems such as low stability due to the swelling of polymeric membranes while pro- cessing high concentrations of CO2 [11,12]. Considering the membrane characteristics and its separation gas transport mechanisms, CMSM
* Corresponding author.
** Corresponding author at: Inorganic Membranes and Membrane Reactors, Sustainable Process Engineering, Chemical Engineering and Chemistry, Eindhoven University of Technology, Eindhoven, the Netherlands.
E-mail addresses: [email protected] (H.R. Godini), [email protected] (F. Gallucci).
Contents lists available at ScienceDirect
Journal of CO2 Utilization
journal homepage: www.elsevier.com/locate/jcou
https://doi.org/10.1016/j.jcou.2022.102378
enables securing a highly selective CO2-permeation in wide range of operating pressures and temperatures [13], which justifies its applica- tion at elevated temperatures and pressures. Developing and tailoring of CMSMs with enhanced CO2 perm-selectivity while being stable at high temperatures and pressures is the focus in making this membrane sep- aration technology viable for steel mill industry [14,15]. CMSMs have shown higher permeability and selectivity values [16] and have been successfully used in high pressure and high temperature applications such as H2/CO2 recovery in steam methane reforming process and CO2/CH4 separation in natural gas (NG) sweetening [17].
CMSMs are synthesized by carbonization of a thermosetting polymer under non-oxidant atmospheres, and depending on the synthesis pro- cedure and the selected precursors, different micro-structures can be tailored in carbon membrane layer [18], which along with their high thermal and chemical stability [19] provide the desired characteristics for wide range of gas separation applications [20]. Facilitating the CO2
adsorption on the tailored CMSM with co-polymerization of Novalac oligomer [20–22] with ethylenediamine has been proven to be an effi- cient method due to increasing the functional groups on the CMSMs microporous structure. However, these membranes have not yet been optimized for high temperatures and pressures applications.
In this research, a CMSM is tailored particularly for this application and its CO2 separation performance under high temperatures and pressures operations assimilating processing the mill gas of a steel plant is experimentally analyzed. Carbon membranes in this work expected to have adsorption diffusion and molecular sieving transport mechanisms which molecular sieving would be ideal for rejection of N2 due to its kinetic diameter (364 pm) compared to H2 (289 pm) and CO2 (330 pm).
However, increasing the adsorption diffusion by enhancing the inter- action of CO2 with nano pores and surface of membrane will result in improving the permselectivity performance of membranes towards CO2 separation. The membranes with functionalized amine groups tested in this research were expected to show an improved perm-selectivity to- wards CO2 when the ethylenediamine content of the deep coating so- lution during their preparation increases. In general, amines adsorb CO2
through chemical adsorption [23] and the presence of these groups on the membrane surface increases the adsorption sites and thereby the selective diffusion of CO2 through the membrane. Moreover, adding ethylenediamine leads to reduction of pore sizes through co-polymerization and branching with Novolac and creating a network of carbon skeleton after carbonization [24]. In the next sections, firstly the membranes development is reported, followed by experimental tests in relevant conditions and the discussion on the obtained results.
2. Experimentation
In this section the details of experimentation including the precursor synthesis, CMSM fabrication steps (dip-dry (polymerize)-carbonize), characterization methods and the performed permeation tests are reported.
2.1. Membrane synthesis and characterization 2.1.1. Synthesis: Materials and method
Formaldehyde (37 % VWR chemicals), Oxalic acid (98 %), ethyl- enediamine for synthesis, N-methyl-2-pyrrolidone (NMP, 99.5 %), and phenol for synthesis were supplied by Merck and used without further purification. CAS # of the used chemicals are summarized in Table 1.
Asymmetric tubular porous alumina with outer diameter (OD) of 14 mm and inner diameter (ID) of 7 mm with an external layer of alumina having 100 nm pore size, were supplied by Inopore GmbH and used as supports.
Porous supports were connected to a dense alumina tube on permeate outlet connection, and on the other side the supports were blocked via glass sealings prepared at 950 ◦C with 10 min curing time as reported before [24]. Firstly, 69 g (0.73 mol) of phenol was melted at 60
◦C in a four neck round bottom glass flask which was equipped with a reflux condenser. Then, 1.5 g of oxalic acid was added to the solution, temperature was risen to 90 ◦C and 52 g (0.64 mol) of formaldehyde solution was added dropwise to the solution. The reaction was carried out for 8 h, the solution was centrifuged at 20 ◦C and was rinsed with denoised water 3 times with a speed of 4000 rpm for 15 min in each step.
The oligomer was collected in a porcelain dish and was vacuum dried at 30 ◦C for 24 h. In the next step, 28 g of the resulted oligomer was dis- solved in 92 g of NMP in a high shear mixer (Thinky ARE-250) with 1800 rpm speed for duration of 30 min in two-stages of mixing. Then, 0.3 g of oxalic acid was added to the solution and mixed for 20 min at 2000 rpm. In the next step, 1.8 g of formaldehyde solution was added to the solution and mixed for two cycles of 30 min at 1500 rpm. Depending on the dipping solution, 0.4 or 1.2 wt percentage (wt%) of ethylenedi- amine was added to the solution and mixed for 30 min at 1000 rpm. The supported membranes were prepared by the dip-coating, of the support in a dipping solution, dry in an oven a temperature and carbonization (Fig. A1) method as we reported before [25].
2.1.2. Membranes characterization
The fabricated CMSMs were characterized for their pore size distri- bution using a perm-porosimeter, selective layer thickness via scanning electron microscopy (SEM, 3D Quanta 250), organic elements using CHNO element analyzer and surface analysis via 3D laser confocal microscopy.
2.1.2.1. Pore size distribution in CMSMS. Measurement of PSD in the fabricated membranes were performed via a custom-made perm- porosimetry machine. Before each test, CMSMs were dried under 2 bar pressure difference between the retentate and permeate at 350 ◦C for 12 h in N2 atmosphere. The details of the setup could be found in our previous report (supplementary information) [26].
2.1.2.2. CHNO analysis. The elemental composition was analyzed in the unsupported membranes using a Thermo Scientific, Flash smart- CHNS/O analyzer. The unsupported samples were prepared by pouring 5 ml of dipping solutions in Teflon dishes and polymerizing them in the oven under Argon atmosphere for 24 h at 90 ◦C and carbonized at the same conditions as supported CMSMs.
2.1.2.3. 3D laser confocal microscopy. To analyze the surface of the fabricated CMSMs, a 3D laser confocal microscope (VKX-3000, Keyence) was utilized with 50× and 150× magnification lenses for determining the quality of surface such as uniformity (surface roughness), existence of cracks and effect of ethylenediamine on the surface quality. All the CMSMs, were tested before and after permeation tests to validate the stability of the membranes in the permeation test conditions.
2.2. Membrane permeation results
Permeation of the CMSMs were tested under wide range of operating conditions for a single component (H2, CO2, and N2) and their mixtures as a function of temperature (50–300 ◦C) and pressure (5–30 bar). The ideal perm-selectivities were calculated according to the ratio of the permeances and were compared with Robeson’s upper bound [27] as a Table 1
CAS # and purity of the used chemicals which are used in synthesis of Novolac precursor and fabrication of CMSMs.
Chemical CAS # Purity (wt%)
Formaldehyde 50–00–0 37
Oxalic acid 144–62–7 98
Ethylenediamine 107–15–3 99.9
N-methyl-2-pyrrolidone 872–50–4 99.5
Phenol 108–95–2 99
performance indicator for the fabricated CMSMs.
The steel plant mill-gas composition was simulated with a three components gas mixture. The conceptual block-flow diagram of the permeation setup is provided in Fig. 1. A schematic representation of a typical carbon membrane is also provided in this Figure.
The represented feed composition with a total flow of up to 50 NL/
min is the potential CO2-rich stream coming from steel industrial plant in which the mole fractions of the dry components are 4 % H2, 22.4 CO2, 73.6 % N2). The gas compositions of the permeate stream and the retentate streams were continually measured using Agilent Micro GC 490. In the mixture, water always was present via a humidifier in saturation conditions which it was also reflecting the effects of possible physio adsorbed intrapore H2O molecules.
3. Results and discussion
The synthesized CMSMs were characterized and tested with regard to their permeation towards different components. The observed results of characterizations and the permeation tests in term of permeation and permeance under each set of experimental conditions were calculated and reported in this section.
3.1. Membrane characterization results 3.1.1. CHNO analysis
Powders of the support-free CMSMs and Novolac oligomer were analysed to determine their elemental composition. The CHNO composition of the non-supported CMSM determined by elemental microanalysis and the calculated number of atoms and molecular weight (MW) of a hypothetical carbonized unit are listed in Table 2. The mo- lecular weight of a novolac structure containing 3 phenolic units is 306, close to the values from microanalysis. For the membrane carbonized without the addition of ethylenediamine, the structure could correspond to a compound with 4 benzene bonded by CH2 with a MW of 380, close to that was calculated from the analysis. The addition of ethylenedi- amine produces an increase of the molecular weight, 4 and 6.4 times for E0.4 and E1.2 respectively. These results indicate that ethylenediamine acts as crosslinker. It was reported that the reaction between dihydroxy biphenyl (DBP), ethylenediamine and formaldehyde in acidic media, polymerization occurs where ethylenediamine is linking to phenolic groups (Ph-CH2-NH-CH2-CH2NH-CH2-Ph) [28]. The ethylenediamine could induce the formation of linear polymers, which is reflected in the reduction of the roughness of the membrane with the addition of eth- ylenediamine (Fig. 2 and Table 3). This also could produce more compact membranes (reduction of the thickness).
As indicated by these micro analysis results, appearance of nitrogen atoms and an increase in the percentages of hydrogen and oxygen atoms
are observed after carbonization. The source of oxygen in the CMSMs is the existing oxygen in the Novolac precursor; introducing ethylenedi- amine to the dipping solution prior to carbonization, increased con- centration of oxygen and hydrogen in the CMSMs. According to Fig. A2, oxygen concentration was measured via EDX-SEM mapping and oxygen concentration significantly higher on the surface of CMSMs compared to CHNS analysis results which could be a result of oxidation of the surface of CMSMs after permselectivity tests. Observing the nitrogen atoms confirm that nitrogen containing functional groups such as amines are existing in the pores and surface of the functionalized CMSMs. The further validation of the effect of functional groups will be performed by analysis of CO2 perm-selectivity tests.
3.1.2. 3D laser microscopy
3D laser confocal microscopy was utilized to study the roughness of the surface of the CMSM of the membrane without the addition of eth- ylenediamine (E0), with the addition of 0.4 (E0.4) and 1.2 (E1.2) wt% of ethylenediamine (H2N-CH2-CH2-NH2) in the dipping solution are shown in Fig. 2; images with higher magnification (150×) are also provided in support information (Fig. A3). It can be observed that the addition of ethylenediamine increases the smoothness of the surface. The rugosity measured with the same equipment is listed in Table 3. the introduction of ethylenediamine and co-polymerization with Novolac, resulted in 20
% and 65 % decrease of average surface roughness (Ra) for membranes E0.4 and E1.2 respectively compared to E0. Furthermore, a decrease of 29 % and 63 % was observed for Rz in membranes E 0.4 and E 1.2 respectively.
The chemical-material mechanism through which adding the ethyl- enediamine reduces the roughness and thickness of the membrane can be explained via co-polymerization of ethylenediamine with Novolac oligomer molecules and creating a branched network of polymer. The new network of the polymer due to its net like structure, will be placed horizontally on the support due to lower energy potential in horizontal position. After carbonization, the network of carbon matrix will be produced from the horizontally placed polymer nets which will result in less thickness and surface roughness in the CMSM. Table 3 reports the average measured roughness indicators of the three membranes.
A decrease of 29 % and 63 % was observed for Rz in membranes E 0.4 and E 1.2 respectively. The reduction of surface roughness and increasing the uniformity of the selective layer will increase the stability of the carbon structure due to reduction of the surface energy at high permeation temperatures. Furthermore, the reduction of surface roughness, results in more hydrophobic surface [29–31] which improves the performance of the membrane in gas feeds containing water vapor such as CO2 separation from post combustion and steel plant off gas process streams.
The impact of adding ethylenediamine therefore can be tracked via
Fig. 1. Block-Flow Diagram of the high-pressure permeation setup with schematic representation of the Molecular Sieve Carbon Membrane.
a) establishing effective adsorption amine sites which facilitates the CO2-permeation primarily through adsorption mechanism, b) better polymer branching and reducing the roughness of the surface and the
thickness of the membrane. These improve both CO2-permeance and CO2-permeability. The measured thickness of the three investigated membranes showed a clear trend in this regard, so that E0, E0.4 and E1.2 Table 2
C,H, O and N composition of novolac and the non-supported CMSM.
C H O N MW (g/mol)
%a # atoms %a # atoms %a # atoms %a # atoms
Novolac 76.97 19 6.79 20.0 16.24 3 0 0 296
E 0 94.12 31 1.53 6.00 4.35 1.07 0 0 395
E 0.4 92.09 122 1.90 30.0 5.05 5.02 0.96 1.1 1590
E 1.2 89.08 187 2.36 59.0 5.74 5.07 2.82 9.04 2519
aSD 0.2%, the number of atoms and molecular weight are calculated from the observed composition.
Fig. 2. The results of 3D surface analysis of CMSMs with 50x magnification lenses with varying ethylenediamine content in the dipping solution.
respectively were measured to have 8.2 µm, 7.5 µm, and 6.7 µm thick- ness (SD 1.9 %, Fig. A4).
3.1.3. Results of pore size distribution (PSD) of CMSMS
The pore size distribution was determined using Helium as inert gas (instead of N2 which is normally used [26] allowing the detection of smaller pores). The calculation of PSD was based on Kelvin’s equation [32]. Fig. 3 summarizes the results of PSD measurements in the inves- tigated CMSMs:
The membrane prepared without ethylenediamine resulted in the biggest pores, with the addition of ethylenediamine, the pore became smaller (E0.4) and for E.12, pores bigger than 0.8 nm are not present.
This behavior is in line with the results obtained above, ethylenediamine produce bigger polymers with more linear structure without big pores present between the polymer ends. This is beneficial for increasing the molecular sieving transport mechanism in separation of CO2 (0.33 nm) from N2 (0.364 nm) and CH4 (0.38 nm). Similarly, the PSD peaks for the membrane with 0.4 % ethylenediamine (E0.4) have been shifted to- wards smaller pores. It should be mentioned that all involved gas components can pass through all investigated membranes in this research. However, the tailored-functionalized membranes intensify the competition between CO2 and other components, which ultimately fa- vors selective CO2-separation due to its higher interactive potentials (due to amine functional groups on the surface as well as the smaller pore size) with the membrane surface and pores.
3.2. Results of permeation tests
Typical observed trends of membrane permeation for single gas and mixtures are reported here, mainly highlighting the counter effects of operating pressure and temperature on the CO2 permeation in the form of permeance (p) and permeability (P) (as formulated in Eqs. 1 and 2):
Pi=Jil
ΔP (1)
pi=Pi
l (2)
Where Pi, pi and Ji, are respectively the permeability, permeance, flux of component i through the membrane, l and ΔP are the thickness of membrane and the applied pressure difference between the retentate and permeate sides.
3.2.1. CMSMs single gas permeations
Fig. 4 demonstrates typical trends of variation of CO2-permeance in wide range of variation of operating pressure and temperature for the three investigated membranes.
As seen in this figure, the CO2-permeance for the E1.2 tailored membrane increases by increasing the pressure up to 20 bar mainly due to the proportional effect of pressure on improving the adsorption over the reachable numerous membrane adsorption sites in this case.
Therefore, 20 bar was considered as an operating pressure, that invest- ing energy to go beyond which cannot be easily justifiable for improving the single gas CO2-permeance indicator.
The observed proportional increase on the CO2-permeance by increasing the temperature is also more pronounced for the E1.2 mem- brane, but can be tracked also for other membranes. This implies that the molecular sieving is the rate determining CO2-transport mechanism with increasing the ethylenediamine content in the dipping solution of the CMSMs. Therefore, applying high temperature will improve the CO2- permeation as it improves the effective movements of molecules through the pores. However, increasing the pressure will lead in an improvement of CO2 surface adsorption and ultimately its permeation when enough adsorption sites are available on the membrane surface.
Fig. 5 shows typical observed trends of CO2-permeability for these membranes while varying temperature and pressure.
As seen here, membrane E1.2 has shown the highest permeability. As explained earlier, the CO2-permeation across the membrane function- alized with ethylenediamine is improved because of the improved contribution of the surface adsorption mechanism and also because of the resulted lower thickness and surface roughness of the functionalized membrane. Later effect can be better visualized by comparing the relative difference between the observed permeance and permeability and their trends in Figs. 4 and 5 for the three investigated membranes (see Eqs. 1 and 2).
3.2.2. CMSMs CO2/N2 ideal gas selectivities
Fig. 6 shows the trends of CO2/N2 ideal selectivity while varying operating temperature. It is seen that the relative permeation across the membranes with higher intensities of amine-functionalized sites, are more sensitive to the variation of temperature as it significantly affects the adsorption rate. As seen in Fig. 6, the highest values of CO2/N2 ideal selectivity are observed in the temperature range (100–150◦C). This is because by increasing the temperature, the permeability for both ni- trogen and carbon dioxide increases. However, in higher temperatures the adsorption diffusion of CO2 will be hampered and therefore reduces the ideal selectivity. This is more pronounced for the E1.2 membrane, over which the intensity of the functionalized group intensifies the contribution of the adsorption separation mechanism.
3.2.3. Perm-selectivities in processing a gas mixture
In order to have a more realistic picture of the separation behavior of the membranes, their separation performance was analyzed while pro- cessing the gas mixture in which the main gas species compete for passing through the membranes assimilating the real separation appli- cation. Fig. 7 shows the observed competitive CO2-separation perfor- mance by tracking the CO2 concentration in the permeate site while varying the operating temperature.
Table 3
The measured surface roughness via 3D laser microscopy for the three investi- gated CMSMs.
Membrane Surface roughness (3D laser microscopy)
50× (nm, SD 1.2 %) 150× (nm, SD 2.3%)
Raa Rzb Ra Rz
E 0 286 2021 202 1280
E 0.4 227 1429 69 475
E 1.2 100 734 94 652
athe average surface roughness,
b difference between the tallest (peak) and deepest (valley) in the surface.
Fig. 3. PSD of CMSMs with varying the ethylenediamine content in the dipping solution. Measured at 70◦C and 2 bar pressure difference.
The real selectivity is calculated in this work according to measured gas concentrations in the permeate and feed streams using a gas chro- matograph (Agilent Micro GC 490). Eq. 3, indicates the real selectivity, S, calculation for gas pair i and j:
Si/j=
Cip Cjp Cif Cjf
(3) Where, Cip, Cij, Cif and Cjf are the molar concentrations of i and j in Fig. 4. Trends and the values of the variation of CO2-permeance versus pressure (right), temperature (left).
Fig. 5. Trends and the values of the variation of CO2-permeability versus temperature (right), pressure (left).
Fig. 6. Trends and the values of the observed CO2/N2 ideal selectivity versus
temperature. Fig. 7. Trends and the values of CO2-concentration in the permeate stream
versus operation temperature.
permeate and feed respectively.
The observed ascending trend in CO2 concentration in the permeate side of the functionalized membranes while increasing temperature, is in line with the expected affect of the recognized rate controlling mecha- nism in this case. The competition between the adsorption separation and molecular sieving in the feed gas species (N2, CO2 and H2, Fig. A5 and A6) will result in the optimum value of the enhancement in the CO2
selectivity according to the amount of ethylenediamine utilized in co- polymerization. For instance, the increase in temperature intensifies molecular sieve diffusion, while hampers CO2-adsorption. This is more pronounced for E1.2 membrane, where the contribution of the adsorp- tion separation is significantly affected in higher temperatures, so that relatively higher CO2-permeation using E0.4 membrane have been recorded. Further increase in the ethylenediamine content in the membrane E1.2 results in further shrinkage of the pores according to the PSD measurements and increases in adsorption sites. With increase of temperature, the competition between H2 and CO2 will result in more favorable conditions for H2 transport in the pores due to the molecular sieving and hampering of CO2 transport due to decrease in adsorption of CO2 molecules at elevated temperatures. This is another reason, why membrane E0.4 showed higher CO2 concentration in the permeate compared to membrane E1.2. Membrane E0.4 outperforms E0 and E1.2 due to the balance between the molecular sieving mechanism to exclude the N2 molecules and meanwhile the adsorption sites enhance the CO2
surface diffusion through the membrane.
It should be highlighted that the observed-reported value of CO2- concentration in the permeate side is an indicator for the competition of gases to pass through the membrane. This is in fact an indicator for the resulted permeation selectivity rather than for the absolute capacity of membrane separation. Reviewing the relative scales of the feed, reten- tate and permeate flowrates will help to get a better understanding of this. 50 NL/min feed flowrate was processed in each test, while only up to 1660 Nml/min flowrate (E1.2, 30 bar, 300◦C) was measured in the permeate side which is 30 times smaller feed flowrate. In this case with considering the permeation of CO2, only 13 % of the CO2 in the feed stream was recovered via membrane E1.2. The optimization of mem- brane surface ratio with respect to the feed flowrate was out of the scope of this work, however, in the large-scale plant of this technology a multi- tubular membrane module will be used to account for this and ensure an efficient utilization of surface per volume membrane gas-membrane contact and CO2 separation. Having considered these, the effect of varying the operating pressure on the CO2 concentrations in the permeate stream is shown by Fig. 8. As mentioned, membrane E0.4 outperformed the membranes E1.2 and E0 in establishing the highest CO2 concentration in the permeate side in all operating pressures at
200◦C. The domination of the molecular sieving transport mechanism at 200◦C indicated by the observed independency of CO2 concentration in the permeate from the operational pressure in the membranes E0.4 and E1.2 as shown in Fig. 8. This is in line with the results of PSD measurements explained earlier. Both functionalized membranes are very selective towards CO2 and the E0.4 membrane which has larger pores allows relative faster separation of CO2.
By increasing the temperature, the CO2/N2 real selectivity increases only for the functionalized membranes as seen in Fig. 9. This is due to the interactive impact of surface adsorption and molecular sieving transport mechanisms which facilitates the CO2 adsorption on the sur- face and its transfer across the functionalized membranes. Depending on the operating temperature, there is an optimum surface adsorbed loading in this regard as it is seen that beyond 200◦C, the membrane with the moderate concentration of functionalized groups, yet with larger pore size shows a higher CO2/N2 real selectivity. Therefore, it can be concluded that the increase in real CO2/N2 selectivity depends on the relative concentration of functionalized groups and PSD in the func- tionalized membranes. In this manner, the observed change in the CO2/ N2 real selectivity can be explained by the relative effect of the rise of temperature on the adsorption diffusion (functional groups) and mo- lecular sieving (PSD), which has dominating impact at lower and higher temperatures respectively. In the absence of CO2-sorption facilitating impact (i.e. non-functionalized membrane E0), solely the competitive permeation of N2 and CO2 in molecular sieving, derived through their relative molecular size and movement, is affecting the CO2/N2 real selectivity.
The effect of operational pressure on the observed CO2/N2 real selectivity could be seen in Fig. 10. The decrease in the average pore size and increase in the adsorption sites by the introduction of ethylenedi- amine functionalized groups over the membranes E0.4 and E1.2 resulted in the more rejection of N2 molecules and enhancing of CO2/N2 real selectivity while increasing pressure.
It is seen that the effect of pressure up to 20 bar is in favor of CO2- permeation selectivity for the E1.2 tailored membrane mainly due to dominant adsorption mechanism, while the impact of pressure on the combination of transfer mechanisms in other membranes resulted multiple maximums in the relative selectivity.
3.2.4. CMSMs performance comparison
All in all, comparing the observed performances of the three inves- tigated membranes in this research indicate that functionalizing the membranes have distinguishably improved the carbon dioxide selective separation so that they have performed beyond Robeson upper bound as shown in Fig. 11.
Fig. 8. Trends and the values of the observed CO2-concentration in the
permeate side versus pressure. Fig. 9. Trends and the values of the observed CO2/N2 real selectivity in the permeate side versus temperature.
As seen in this figure, the recorded trends, and the values of CO2/N2
ideal perm-selectivities show the functionalization has a significant improving effect on the selectivity towards more CO2-permeation indi- cating the clear impact of surface adsorption in combination with mo- lecular sieving transport mechanisms.
4. Conclusions
The designed sets of experimentations and the selected fabricated membranes with their established characteristics through the applied synthesis-functionalization, enabled better understanding the func- tioning mechanism of carbon membranes. The performance of CMSMs was examined under wide range of variation of operating temperature and pressure as the main engineering parameters in the operation of such separation system. It was observed that there is an optimum level of functionalization through which the highest selective CO2-separation, beyond Robeson upper bound, can be secured. It could also be concluded that a trade-off between the facilitating impact of enhanced temperature of the permeation of all gas species could be encountered with its undesired impact on the selectivity of CO2-separation explained through the reduction of surface adsorption mechanism. These obser- vations and the conclusion are valuable assets in designing an efficient membrane system for selective CO2-separation from mill gas of
industrial steel plant and eventually even for a membrane reactor application which is the subject of our forthcoming reports. These contribute in improving the environmental and energy efficiency of steel plant. Significant levels of process intensification, energy efficiency and economic benefits as well as emission reduction and environmental impacts are expected using this approach.
CRediT authorship contribution statement
A. Rahimalimamaghani: Conceptualization, Data curation, Formal analysis, Investigation, Writing – original draft. H.R. Godini: Concep- tualization, Writing – original draft. M. Mboussi: Investigation. A.
Pacheco Tanaka: Supervision, Writing – review & editing, Methodol- ogy. M. Llosa Tanco: Supervision, Writing – review & editing, Meth- odology. F. Gallucci: Funding acquisition, Project administration, Resources, Supervision, Writing – review & editing, Methodology.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Data availability
Data will be made available on request.
Acknowledgements
The project has received funding through the NWO P16-10 project.
Appendix A. Supporting information
Supplementary data associated with this article can be found in the online version at doi:10.1016/j.jcou.2022.102378.
References
[1] Greenhouse Gas Emissions - Our World in Data Available online: 〈https://ourworl dindata.org/greenhouse-gas-emissions?utm_source=miragenews&utm_medium
=miragenews&utm_campaign=news〉(accessed on 27 June 2022).
[2] Schl¨ogl, R. et al. Novel Carbon Capture and Utilisation Technologies: Research and Climate Aspects. 2018.
[3] (PDF) Steel, Aluminium and Carbon: Alternative Strategies for Meeting the 2050 Carbon Emission Targets Available online: 〈https://www.researchgate.net/publi cation/242582149_Steel_aluminium_and_carbon_alternative_strategies_for_meetin g_the_2050_carbon_emission_targets〉(accessed on 27 June 2022).
[4] Global Energy Review: CO2 Emissions in 2021 – Analysis - IEA Available online:
〈https://www.iea.org/reports/global-energy-review-co2-emissions-in-2021–2〉
(accessed on 27 June 2022).
[5] R.W. Baker, K. Lokhandwala, Natural gas processing with membranes: an overview, Ind. Eng. Chem. Res 47 (2008) 2109–2121, https://doi.org/10.1021/
IE071083W/ASSET/IMAGES/LARGE/IE071083WF00019.JPEG.
[6] D.C. Rosenfeld, H. B¨ohm, J. Lindorfer, M. Lehner, Scenario analysis of implementing a power-to-gas and biomass gasification system in an integrated steel plant: a techno-economic and environmental study, Renew. Energy 147 (2020) 1511–1524, https://doi.org/10.1016/J.RENENE.2019.09.053.
[7] Decarbonization in Steel | McKinsey Available online: 〈https://www.mckinsey.co m/industries/metals-and-mining/our-insights/decarbonization-challenge-for-st eel〉(accessed on 27 June 2022).
[8] H. Li, K. Haas-Santo, U. Schygulla, R. Dittmeyer, Inorganic microporous membranes for H2 and CO2 separation – review of experimental and modeling progress, Chem. Eng. Sci. 127 (2015) 401–417, https://doi.org/10.1016/J.
CES.2015.01.022.
[9] Z. Yang, W. Guo, S.M. Mahurin, S. Wang, H. Chen, L. Cheng, K. Jie, H.M. Meyer, D.
en Jiang, G. Liu, et al., Surpassing robeson upper limit for CO2/N2 separation with fluorinated carbon molecular sieve membranes, Chem 6 (2020) 631–645, https://
doi.org/10.1016/J.CHEMPR.2019.12.006.
[10] Y. Han, W.S.W. Ho, Polymeric membranes for CO2 separation and capture, J. Membr. Sci. 628 (2021), https://doi.org/10.1016/J.MEMSCI.2021.119244.
[11] G.C. Kapantaidakis, G.H. Koops, M. Wessling, S.P. Kaldis, G.P. Sakellaropulos, CO2 plasticization of polyethersulfone/polyimide gas-separation membranes, AIChE J.
49 (2003) 1702–1711, https://doi.org/10.1002/AIC.690490710.
Fig. 10. Trends and the values of the observed CO2/N2 real selectivity versus, top) temperature, bottom) pressure.
Fig. 11. The observed trends of CO2/N2 ideal selectivity; Robeson upper bound (data correspond to 20 bar operating pressure).
[12] M.S. Suleman, K.K. Lau, Y.F. Yeong, Plasticization and swelling in polymeric membranes in CO2 removal from natural gas, Chem. Eng. Technol. 39 (2016) 1604–1616, https://doi.org/10.1002/CEAT.201500495.
[13] L. Lei, A. Lindbråthen, X. Zhang, E.P. Favvas, M. Sandru, M. Hillestad, X. He, Preparation of carbon molecular sieve membranes with remarkable CO2/CH4 selectivity for high-pressure natural gas sweetening, J. Membr. Sci. 614 (2020), 118529, https://doi.org/10.1016/J.MEMSCI.2020.118529.
[14] M.B. H¨agg, J.A. Lie, A. Lindbråthen, Carbon molecular sieve membranes: a promising alternative for selected industrial applications, Ann. N. Y Acad. Sci. 984 (2003) 329–345, https://doi.org/10.1111/J.1749-6632.2003.TB06010.X.
[15] L. Lei, L. Bai, A. Lindbråthen, F. Pan, X. Zhang, X. He, Carbon membranes for CO2 removal: status and perspectives from materials to processes, Chem. Eng. J. 401 (2020), 126084, https://doi.org/10.1016/J.CEJ.2020.126084.
[16] K.M. Steel, W.J. Koros, Investigation of porosity of carbon materials and related effects on gas separation properties, Carbon N. Y 41 (2003) 253–266, https://doi.
org/10.1016/S0008-6223(02)00309-3.
[17] W.J. Koros, C. Zhang, Materials for next-generation molecularly selective synthetic membranes, Nat. Mater. 16 (2017), https://doi.org/10.1038/nmat4805.
[18] L. Forster, C. D’Agostino, M.A. Llosa-Tanco, V. Spallina, C. Brencio, F. Gallucci, M. Lindley, S.J. Haigh, D.A. Pacheco-Tanaka, Tailoring pore structure and surface chemistry of microporous alumina-carbon molecular sieve membranes (Al-CMSMs) by altering carbonization temperature for optimal gas separation performance: an investigation using low-field NMR relaxation measurements, Chem. Eng. J. 424 (2021), 129313, https://doi.org/10.1016/J.CEJ.2021.129313.
[19] X. Zhang, H. Hu, Y. Zhu, S. Zhu, Carbon molecular sieve membranes derived from phenol formaldehyde novolac resin blended with poly(ethylene glycol), J. Membr.
Sci. 289 (2007), https://doi.org/10.1016/j.memsci.2006.11.047.
[20] G.I. Konishi, T. Tajima, T. Kimura, Y. Tojo, K. Mizuno, Y. Nakamoto, Direct synthesis of functional novolacs and their polymer reactions, Polym. J. 42 (2010) 443–449, https://doi.org/10.1038/pj.2010.26.
[21] G.I. Konishi, New class of aromatic polymers prepared by addition-condensation of designed phenol with formaldehyde, Kobunshi Ronbunshu 66 (2009) 331–339, https://doi.org/10.1295/koron.66.331.
[22] G.I. Konishi, Precision polymerization of designed phenol: new aspects of phenolic resin chemistry, J. Syn. Org. Chem. Jpn. 66 (2008) 705–713, https://doi.org/
10.5059/yukigoseikyokaishi.66.705.
[23] Y. Wu, J. Xu, K. Mumford, G.W. Stevens, W. Fei, Y. Wang, Recent advances in carbon dioxide capture and utilization with amines and ionic liquids, Green Chem.
Eng. 1 (2020), https://doi.org/10.1016/j.gce.2020.09.005.
[24] M.A. Llosa Tanco, D.A. Pacheco Tanaka, S.C. Rodrigues, M. Texeira, A. Mendes, Composite-alumina-carbon molecular sieve membranes prepared from novolac resin and boehmite. Part I: preparation, characterization and gas permeation studies, Int. J. Hydrog. Energy 40 (2015) 5653–5663, https://doi.org/10.1016/J.
IJHYDENE.2015.02.112.
[25] A. Rahimalimamaghani, D.A. Pacheco Tanaka, M.A. Llosa Tanco, F. Neira D’Angelo, F. Gallucci, Effect of aluminium acetyl acetonate on the hydrogen and nitrogen permeation of carbon molecular sieves membranes, Int. J. Hydrog. Energy 47 (2022) 14570–14579, https://doi.org/10.1016/J.IJHYDENE.2022.02.198.
[26] A. Rahimalimamaghani, D.A. Pacheco Tanaka, M.A. Llosa Tanco, F. Neira D’Angelo, F. Gallucci, Effect of aluminium acetyl acetonate on the hydrogen and nitrogen permeation of carbon molecular sieves membranes, Int J. Hydrog. Energy 47 (2022) 14570–14579, https://doi.org/10.1016/J.IJHYDENE.2022.02.198.
[27] L.M. Robeson, The upper bound revisited, J. Membr. Sci. 320 (2008) 390–400, https://doi.org/10.1016/J.MEMSCI.2008.04.030.
[28] K. Soueidan, S. Chen, H.R. Dajani, al -, Y. Zheng, C. Wang, F. Yang, S.P. Chakole, K.
A. Nandekar, W.B. Gurnule, Photoluminescent studies of 2, 2’-dihydroxybiphenyl, ethylenediamine - formaldehyde copolymer, J. Phys. Conf. Ser. 1913 (2021), 012062, https://doi.org/10.1088/1742-6596/1913/1/012062.
[29] M.F. Ismail, B. Khorshidi, M. Sadrzadeh, New insights into the impact of nanoscale surface heterogeneity on the wettability of polymeric membranes, J. Membr. Sci.
590 (2019), 117270, https://doi.org/10.1016/J.MEMSCI.2019.117270.
[30] W. Qing, J. Wang, X. Ma, Z. Yao, Y. Feng, X. Shi, F. Liu, P. Wang, C.Y. Tang, One- step tailoring surface roughness and surface chemistry to prepare
superhydrophobic polyvinylidene fluoride (PVDF) membranes for enhanced membrane distillation performances, J. Colloid Interface Sci. 553 (2019) 99–107, https://doi.org/10.1016/J.JCIS.2019.06.011.
[31] B. He, J. Lee, Dynamic wettability switching by surface roughness effect, Proc.
IEEE Micro Electro Mech. Syst. (MEMS) (2003) 120–123, https://doi.org/10.1109/
MEMSYS.2003.1189702.
[32] T. Tsuru, T. Hino, T. Yoshioka, M. Asaeda, Permporometry characterization of microporous ceramic membranes, J. Membr. Sci. 186 (2001) 257–265, https://doi.
org/10.1016/S0376-7388(00)00692-X.